Project Summary Current implantable cardioverter defibrillator (ICD) implant rates are approximately 150,000/year based on registry data; however, many patients meeting standard implantation criteria for ICDs to prevent sudden death from an initial episode of sustained ventricular tachycardia (VT) or ventricular fibrillation (VF) do not benefit. In fact, most ICDs are now implanted exactly this indication, which is usually described as primary prevention of sudden death. The consequences of implanting ICDs in patients unlikely to benefit is compounded by the fact that multiple replacement devices may be needed over a patient's lifetime because of eventual battery depletion. In fact, over 40% of the ICDs implanted each year are replacement devices for patients with prior ICDs. Unfortunately, these procedures are associated with a significant risk of complications such as device infections, which could result in net harm to patients unlikely to receive beneficial therapies for ventricular arrhythmias over a lifetime. In patients who are also not dependent on pacing for bradycardia or heart failure and unlikely to experience improved survival with the ICD, either turning off the device or explanting the device is feasible and may be preferred by many. Considering then the potential harm from unnecessary procedures and the already unsustainable healthcare expenditures on ICDs, there is a pressing and unmet need to identify patients who may do as well without ICD generator replacement when battery depletion occurs because of a low risk of sudden death at this time. In order to address this unmet need, we will analyze outcomes after hundreds of thousands of replacement and new ICDs in Version 2 of the ICD Registry maintained by the National Cardiovascular Data Registry (NCDR) and compare them to outcomes in patients with heart failure without ICDs. The two aims will provide important data regarding two key issues of great interest to the field. The first aim it to determine if having a significant improvement in left ventricular (LV) function at the time of the replacement ICD procedure is associated with a decreased relative and absolute ICD survival benefit. By virtue of having greater LV function at this time based on an LV ejection fraction over 35%, patients no longer technically meet one of main criteria used to determine who should receive an ICD for primary prevention of sudden death. The second aim is to determine how the risk of sudden death and the expected survival benefit changes from the time of the initial ICD implantation to the time of ICD replacement. In both aims, we will apply two advanced prediction models (Seattle Proportional Risk Model [SPRM] and Seattle Heart Failure Model [SHFM]) to test our hypothesis that patients with a low predicted risk of sudden death at the time they undergo ICD replacement are unlikely to have improved survival from a replacement ICD. In summary, we expect this project to promote precision medicine, good patient care, and efficient allocation of healthcare resources by demonstrating how we can provide better risk stratification for patients who are candidates for replacement ICDs for primary prevention of sudden death.